Lessons from the 1979-82 Monetary Policy Experiment
Milton Friedman
The American Economic Review, Vol. 74, No. 2, Papers and Proceedings of the Ninety-Sixth
Annual Meeting of the American Economic Association. (May, 1984), pp. 397-400.
Stable URL:
http://links.jstor.org/sici?sici=0002-8282%28198405%2974%3A2%3C397%3ALFT1MP%3E2.0.CO%3B2-J
The American Economic Review is currently published by American Economic Association.
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/about/terms.html. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained
prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in
the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/journals/aea.html.
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.
The JSTOR Archive is a trusted digital repository providing for long-term preservation and access to leading academic
journals and scholarly literature from around the world. The Archive is supported by libraries, scholarly societies, publishers,
and foundations. It is an initiative of JSTOR, a not-for-profit organization with a mission to help the scholarly community take
advantage of advances in technology. For more information regarding JSTOR, please contact [email protected].
http://www.jstor.org
Sat Feb 9 15:51:22 2008
Lessons from the 1979-82 Monetary Policy Experiment A "monetary policy experiment" was conducted from October 1979 to the summer of
1982, but many commentators have misinterpreted its character and have drawn the
wrong conclusions from the evidence it generated. They have termed it a "monetarist
policy" and have interpreted the results as a
failure of monetarism.
Though the Federal Reserve System's rhetoric was "monetarist," the actual policy that
it followed was antimonetarist (see my 1983
article). And the evidence generated by the
experiment strongly supports rather than
contradicts the propositions that recommend
a monetarist policy.
In its October 6, 1979 announcement that
initiated the experiment, the Fed described
its changed procedures as designed "to support the objective of containing growth in
the monetary aggregates." That is indeed a
monetarist objective. However, a monetarist
policy involves not only targeting monetary
aggregates, but also-as a major and central
element-achieving a steady and predictable
rate of growth in whatever monetary aggregate is targeted. By this essential criterion,
the experiment was antimonetarist: as Table
1 shows, the volatility of monetary growth
during the experiment was about three times
as high as earlier. Indeed, monetary volatility
was higher during the three years of the
experiment than in any earlier three-year
period since at least the end of World War
11.' The purely rhetorical character of mone-
*Senior Research Fellow, Hoover Institution, Stanford, CA 94305.
'The failure of the Fed to follow a monetarist policy
is not surprising. If at the outset of the experiment, each
member of the Board of Governors had been asked,
"Are you now, or have you ever been, a monetarist?,"
not a single one would have answered "yes." In a
column commenting on the October 6 , 1979 announcement of the change in policy, I wrote that " those of us
who have long favored such a change have repeatedly
licked our wounds when we mistakenly interpreted
earlier Fed statements as portending a change in operat-
Ten Quarters:
M1
Adjusted
Monetary
Base
Prior to October 1979
After October 1979
1.59
5.64
.94
2.71
tary targeting is clear also from the Fed's
failure to hit its targets. M1 was outside the
target range in six out of ten quarters, and
t h s despite the generous width of the range
between the lower and upper growth rate
targets plus an annual shfting of the base to
which the target growth rates were applied.
Many monetarists believe that slowing
monetary growth will reduce inflation more
promptly and at a smaller cost in terms of
reduced output and higher inflation if it is
announced in advance than if the slowing is
not anticipated. However, their belief depends on the pre-announced slowing of
monetary growth being widely believed by
the relevant economic agents. Such belief
was not widely present, even in October 1979,
when the new policy was first announced.
And the wide gyrations in monetary growth
rates in subsequent months rapidly disillusioned any naive agents who initially accepted the Fed's rhetoric as a guarantee of
steady and predictable monetary growth.
Since there was no credible pre-announced
slowing, it could not have had any of the
effects attributed to such a slowing by some
monetarists. Similarly, the failure of the
policy to achieve such effects cannot be regarded as contradicting monetarist predictions.
ing procedures. I hope this time will be different-but
remain skeptical until performance matches pronouncements" (1979, p. 39).
398
AEA PAPERS A N D PROCEEDINGS
The real experiment was in the operating
procedures adopted-the
use of nonborrowed reserves as the instrument. In effect,
that meant a reversion to the "free reserves"
approach of the 1920's. Combined with
lagged reserve requirements, the new approach produced enhanced volatility in
monetary growth, and, as a consequence, in
both interest rates and economic activity. As
a citizen, I deplore the results, which, as I
have argued elsewhere (1983), included imposing a much higher cost for the achieved
reduction of inflation than was necessary. As
a scientist, I am delighted to have the unintentional experimental evidence on the effect
of sharp swings over short periods in the rate
of monetary growth.
These sharp swings provide some evidence
on two monetarist propositions: first, that
different monetary aggregates move together;
second, that movements in monetary aggregates produce corresponding movements in
nominal income.
I. Different Monetary Aggregates
In judging the evidence on claims by
monetarists that different monetary aggregates move together, it is important not to be
confused by labels. The aggregates as currently defined do not correspond to the aggregates with respect to which those claims
were made. They were made primarily with
respect to M 1 and M2 as then defined. The
current aggregate labelled M 2 is a much
broader aggregate than the earlier M2. Indeed it is nearly identical, both conceptually
and numerically, with the aggregate that
Anna Schwartz and I labelled M4 in our
Monetary ~ t a t i s t i c sThe
. ~ current M1 is conceptually, though in the earlier years not
numerically, closer to the aggregate we
labelled M 2 rather than to our M I because,
2 ~ oexample,
r
for January 1960, our estimate of M 4
is $297.4 billion; the Federal Reserve Board's estimate
of the current M 2 , $298.2 billion; our estimate of our
M 2 , $208.9 billion (see our study, 1970, p. 47). For
January 1983, I estimate the counterpart to the earlier
M 2 to have been $1012.3 billion. whereas the Fed's
estimate of the current M 2 is $2176.1 billion, or more
than twice as large.
M A Y 1984
like our M2, it includes deposits bearing
i n t e r e ~ t The
. ~ current aggregate that is conceptually closest to the earlier M I is the
monetary base.
Few if any monetarists ever recommended
the use of such broad aggregates as the current M 2 or M3 as monetary targets-certainly, this one did not. The closest current
approximations to the aggregates they recommended are therefore the current M I and
the monetary base.4
For the five years, 1978 to 1982, the correlation between the fourth-quarter to fourthquarter rates of growth of M1 and the base
adjusted for reserve requirement changes is
.89; of M I and the unadjusted base, .63. On
the other hand, the correlation of M I and
M 2 is .34; of M I and M3, .17; of M 2 and
M3, .17.
11. Money and Income
The asserted relation between movements
in a monetary aggregate and nominal income
that is relevant to current policy is about
cyclical effects. In judging this proposition, a
year is too long a time unit to use, especially
for the period from 1980 to 1983, since it has
been characterized by a succession of abnormally short cyclical phases: a six-month contraction in 1980, followed by a twelve-month
expansion and then a sixteen-month contraction, interrupted by a one-quarter revival.
Monetarists attribute this result to the corresponding abnormally short and exceptionally
volatile gyrations in monetary growth. Moreover, monetarists have typically concluded
that the lag of the rate of change of nominal
income behind the rate of change of M1 is
3 ~ nthe earlier years, before the introduction of
checkable deposits that paid interest, the current M 1
was numerically the same as our M 1 , which included
only non-interest-bearing assets. For example, for
January 1960, the Federal Reserve Board's estimate of
the current M 1 is $141 billion; our estimate of M 1 is
$141.7 billion (Friedman-Schwartz, p. 47). However, for
January 1983, I estimate the closest counterpart to the
earlier M 1 to have been $373.7 billion, whereas the
Fed's estimate of the current M 1 is $482.1 billion.
4 0 r a reconstruction of the equivalent of the earlier
M 2 , something that I have not attempted.
V O L 74 NO. ?
MONETARISM: LESSONS FROM THE POST-1979 EXPERIMENT
399
TABLE
2-SWINGS IN .v1 AND NOMINAL
GNP
ONEQUARTER
LATER, QUARTERLY
DATA,
1979:l TO 1983:4
Annual Rarc of Ch'lnge
Per~od
for ,MI
Nuniher
of
Quarters
4
Z
4
Norr: Periods for: MI, 79:3-R3:3; G N P , 79:4-83:4
Monctar)
Ra\e
X?
I
64
7X
33
1
2
4
96
73
9.3
MI
74
15
10 1
3.2
11 0
47
12.6
(;,VP One
Quartcr
Latcr
102
5.2
13 9
11
66
2.6
10 3
Pcriod
forGhP
79:l to P0:l
80.1 lo X0:3
XO-? lo X1:3
X1:3 lo 82.1
X2:l lo 82.2
X2:2 lo 82:4
82 4 to 83.4
thereby to speed up and render more consistent the reaction.5
The close relation between money and
nominal income is brought out in a different
way in Table 2, which distinguishes the
successive periods of rapid and slow growth
in M I . The one-to-one relation between ups
and downs in MI growth and in GNP growth
one quarter later is striking. There is a sirnilar one-to-one relation between ups and
downs in M 1 and in the monetary base.
111. Money and Inflation
about six months on the average, in general
ranging from three to nine months.
Figure 1 plots the quarter-to-quarter rates
of change of M1 and of GNP one quarter
later for the period of the "monetary policy
experiment." From 1981 on, the relation
is extraordinarily close-indeed, instead of
being less close than during the earlier years,
it is considerably closer. For 1980, there are
significant discrepancies, almost surely attributable to President Carter's imposition
and subsequent removal of credit controls.
The correlation between the two series is .46
for the period as a whole; .71, eliminating
the quarters affected by the credit controls.
Two things are notable about the relation
between money and income in these years:
first, the lag is both shorter on the average
and less variable than in earlier years, second, the relation is unusually close. I believe
that both are a consequence of the exceptionally large fluctuations in M1 growth. The
effect was to enhance the importance of the
monetary changes relative to the numerous
other factors affecting nominal income and
The long-period evidence suggests that inflation has much inertia and that the lag
between money and inflation is of the order
of two years. Table 3 shows that this relation
has held in recent years as well. There is a
one-to-one relation between movements in
monetary growth, and in the GNP deflator
two years later over successive two-year periods since 1971. The decline in inflation from
1979-81 to 1981-83 was decidedly larger
than might have been expected from the
decline in monetary growth. I attribute that
not to a pre-announced slowing of monetary
growth, but rather to the exceptional volatility of monetary growth, whch increased the
degree of perceived uncertainty and thereby
increased the demand for money.
'For longer periods, annual data do provide relevant
evidence on the relation between M1 and nominal income. For the 111 years from 1871 to 1981, the contemporaneous correlation between rates of change of M1
and nominal income is .72; for the 24 years from 1960
to 1983, it is .75, and rises to .86 if the G.VP data are for
a year ending one quarter later than the M1 data.
400
M A Y 1984
A E A PAPERS A N D PROCEEDINGS
Again, I attribute t h s result to the larger
amplitude and shorter duration of the recent
swings in monetary growth.
Annual Rate of Change Over Eight Quarters Period for
Money
M1
Deflator Eight
Quarters Later
Period for
Deflator
The increased rate of monetary growth in
the 1981-83 biennium suggests that we have
passed the trough in inflation and that inflation will be decidedly higher from 1983 to
1985 than it was from 1981 to 1983.
IV. Money and Real Economic Growth
The inertia in inflation and the lengthy lag
between monetary change and inflation
mean, of course, that the short-run influence
of money on nominal income will be reflected primarily in movements in real income. And that is the case for the period
under consideration. The correlation between quarter-to-quarter rates of change of
M1 and real GNP one quarter later is .54 for
the whole period covered by Figure lt and
.86 excluding the quarters affected by credit
controls. These correlations are even hgher
than those for nominal income, though that
has generally not been the case in the past.
V. Conclusion
The evidence generated by the misinterpreted monetary policy experiment of 1979
to 1982 is entirely consistent with the empirical conclusions about the relation between
money, income, and prices that monetarists
have drawn from earlier evidence. If anything, the recent evidence strengthens the
case for a policy of steady predictable monetary growth. Of course, evidence for three
years and a single country cannot add much
to evidence provided by studies covering
more than a century of experience and many
countries. But the particular three years do
contribute disproportionately precisely because the policy actually followed deviated
so widely from the policy recommended by
monetarists.
REFERENCES
Friedman, Milton, "Monetarism in Rhetoric
and in Practice," Bank of Japan Monetary
and Economic Studies, October 1983, 1,
1-14.
-*
"Has the Fed Changed Course?,"
Newsweek, October 22, 1979, p. 39.
and Schwartz, Anna J., Monetary Statistics of the United States, New York:
Columbia University Press for the National Bureau of Economic Research, 1970.